SOFTWARE PACKAGES
|
DIAMOND SoftwareDIAMOND is a MATLAB software toolbox which implements several state-of-the-art techniques for experimental modal analysis, damage identification, and finite element model correlation and refinement in a menu and button-driven graphical user interface. Availability and PricingCurrently, we are exploring with our employer the procedures for distributing DIAMOND to the public. The most likely scenario is the distribution as a third-party MATLAB toolbox, with a nominal site license fee (possibly around $500). This would require the user to purchase MATLAB version 5.1 and the MATLAB Signal Processing toolbox separately. Currently, we are planning to support UNIX and Windows platforms. Damage IdentificationThe damage identification capabilities in DIAMOND will consist initially of two algorithms: The first is strain energy analysis, which examines changes in the estimated strain energy in user-defined Beam or Plate elements based on the curvature of the undamaged and damaged mode shapes. The second algorithm is flexibility analysis, which uses the changes in the measured flexibility matrix to assess potentially damaged areas in the structure. Finite Element Model UpdatingThe initial release of DIAMOND will contain a rudimentary capability for Finite Element Model refinement using the Minimum Rank Perturbation technique. Future versions of the software will extend this capability to incorporate a variety of model refinement, model reduction / mode shape expansion, and statistical parameter updating techniques. Modal AnalysisThe initial release of DIAMOND will have the capability of analyzing modal data from single or multi-reference experiments to determine modal frequencies, modal damping, and mode shapes. Several commonly used modal curve-fitting algorithms are currently implemented, including operating deflection shape fitting, rational polynomial frequency domain fitting, complex exponential time domain fitting, and eigensystem realization algorithm (ERA) time domain fitting. DIAMOND also has the capability to estimate statistical confidence intervals on the identified modal parameters. These confidence intervals allow the analyst to state 95% certainty limits on the modal parameters. For example, "The first modal frequency of the structure is 11.35 Hz with a 95% confidence interval between 11.25 Hz and 11.45 Hz," indicates that when the modal frequency is measured under the conditions of that test 100 times, it will be expected to fall into the range 11.25 Hz - 11.45 Hz for 95 of those measurements. The statistical confidence intervals are computed using two different techniques: The first is a Monte Carlo simulation-based technique, which uses confidence intervals on the averaged frequency response function data (computed using the measured coherence function) to compute corresponding confidence intervals on the identified modal parameters. The second technique uses individual (non-averaged) frequency response functions to compute the confidence intervals on the identified parameters using a Bootstrap analysis approach.
|
Related Links |